Re: Previous "will not respond to apneas above 10cm" threads

[mymontreal]

There have been many threads where the subject of "machines not responding to Apneas above 10cm" has been discussed.

Snoredog, I know you have clarified a number of times (at least in the case of the Remstar APAP machine) that the pressure will not increase above 10.0cm in order to respond to a Frank Apnea & that it will only go above 10.0cm to deal with Flow Limitation or Snoring events. (from what I have read, the only way around this is to increase the Min Auto setting to a value > 10.0)

Apologies, but I am still very confused on this and trying to figure out what my data represents - I have attached detailed results from 3 different days below - (my Min Auto setting is 10, my Max Auto setting is 15 & I am using a ramp that starts at 8.0) - in each case my indicated 90% AutoCPAP is between 11 and 15 cm, yet I see nothing/virtually nothing in the way of Flow Limitation or Snoring events to explain why the pressure is rising above 10.0, if it is not to deal with Apnea Events... what am I missing, or is my data screwed up because I still have leak issues to deal with ?

Thanks
Mike

[dsm]

Mike,

When Snoredog talks of machines not responding to apneas above 10CMs he is usually referring to the Resmed machines that use the A10 algorithm.

The machines do respond to apneas over 10CMs but that then turns into a debate of what apnea means. If we accept that there can be apneas that are 100% obstructive (no air flow) and that there are apneas that are called flow-limitations/hypopneas (some airflow) and there are apneas that are Central (no airflow) then the A10 algorithm can be correctly said to deal with some of these apneas, over 10CMs, but it does so selectively.

The A10 algorithm will regard certain of these apneas as potential Centrals and on that basis will reasonably seek leave them alone. Continually slowly adding pressure to a central apnea compounds it if the pressure keeps slowly rising (which is waht AUTOs do). If we look at how a timed bilevel reponds to a Central, we can see that it applies a burst of pressure (flips from epap to ipap) for a short period in the hope that that ipap burst jolts the sleeper back into breathing. Slowly increasing pressure doesn't clear a central but can disturb the sleeper if the pressure rise isn't stopped.

If A10 detects snores over 10 CMs it will rightly regard these as precursor events leading to a genuine obstructive apnea so it *does* raise its pressure & thus is responding to a *looming* obstructive apnea. If there are no snores and suddenly there is no flow (which could be a Central or an obstructive apnea) the A10 algorithm quite reasonably decides to leave it to nature. So, based on the situation that the A10 algorithm did not detect the typical obstructive apnea precursor events (snores & flow-lims/hypopneas) the correct statement is

The A10 algorithm won't respond to Obstructive apneas over 10 CMs *if* there are no obstructive apnea pre-cursor events that precede the obstruction. And, in reality there is nearly always some pre-cursor obstructive apnea events that precede obstructive apneas.

I believe myself, that it is a mis-statement to say that 'the A10 algorithm doesn't respond to 'apneas' over 10 CMs' without spelling out clearly what is meant by apnea- because not clarifying that meaning twists the reality.

DSM

[Blue Snuffle nose]

Hello there,

I am above 10cm Ho in pressure as well.......I am 12 cm ho in pressure in my cpap and it responds well to it. However, I have a older model cpap, S6 Elite Sullivan cpap machine which I have been hiring for a few years. I haven't got a moden cpap as I can't afford buying one, $1000 + dollers for a more moden model.

I guess my lesson is you can't always find everything you want sometime in the lastest technology......{not knocking newer cpaps] As I have used the older cpap and years and I am happy with it. I used to use the alarm feature, which shut the cpap off if your mask is not or connected to it. That failed on me so I had get the cpap people to help me shut off the feature.

Now I just use "plain cpap".

[ozij]

1. The issue of responding after an apnea is identified is only relevant to automatic pressure machines. On fixed pressure machines, the pressure is fixed and if a person needs a pressure of 12 to keep them from having most of their apnea, that's what they'll have. Their pressure is independent of events.

2. Automatic machines have to distinguish obstructive apneas - which justify a rise in pressure - from non-obstructives ones, that do not need raise in pressure.

3. The companies who make automatic machines go about this identification process in different ways.

Resmed's Auto Set Algorithm is described here: http://www.resmed.com/en-au/clinicians/ ... clinicians

And here: http://www.talkaboutsleep.com/sleep-dis ... chat.htm#6

The A10 algorithm increases pressure in response to Flow Limitation, Snore, and Apnea up to 10cm H2O. Above 10cm H2O, pressure response to Flow Limitation and Snore continues, but there is no response to Apnea. AutoSet Spirit and AutoSet T do not differentiate between obstructive and central apneas. Increasing pressures above 10cm H2O in response to apnea can lead to "runaway" central apneas,

This algorithm is still in use in the S8 machines.
If a person has apneas, but no snores or flow limitations above a 10 cwm, then a Resmed may not be the best machine for treatment. However, the majority of people have flow limitation and snores as well as apneas, and Resmed think it is safe enough to stop at 10.

Respironics identity apneas /hypopneas that do not respond to pressure, and then drop the pressure back down after having raised it too much.

Above 8 cmH2O pressure, the pressure increase for sustained apneas/hypopneas is limited to 3 cmH2O above the pressure setting at the onset of the apnea/hypopnea sequence. The pressure setting at the onset of the sequence is called the "Onset Pressure." The pressure 3 cmH2O above the "Onset Pressure" is called the "NRAH Threshold." "NRAH" is an acronym for "Non-Responsive Apnea/Hypopnea."
<snip>
When the device encounters a non-responsive apnea or hypopnea, it will decrease pressure by 2 cmH2O and hold the pressure for 15 minutes. During this 15-minute "NRAH-hold" period, the pressure will be changed only in response to detection of a sequence of vibratory snore events. The reason for this increase in pressure followed by a decrease in pressure when an apnea/hypopnea is not responsive is to allow the device to respond appropriately to an event that is not treatable by increases in pressure, such as a central apnea

.
Emphasis added by me.

Your machine is raising the pressure, Mike, because your apneas are responsive to those pressure raises. You do not not sustained apneas that are not responding the rise in pressure.

On Puritan Bennett automatic machines, you can tell the machine what the maximum pressure will be in response to apnea - the default is 10, but it can be set both higher or lower. The machine is also capable of identifying heartbeat oscillations in the air flow. Their existence indicates that the airway is open, and therefore the machine will not raise pressure in response to apneas accompanied by cardiogenic oscillations - no matter what pressure they occur at.

All of these algorithms will raise pressure as higher than 10 if they run into flow limitation or snores.

O.

Edited some typos....

[rested gal]

Great explanation, ozij.

[jnk]

Wow! Bookmarked!

[mymontreal]

1. The issue of responding after an apnea is identified is only relevant to automatic pressure machines. On fixed pressure machines, the pressure is fixed and if a person needs a pressure of 12 to keep them from having most of their apnea, that's what they'll have. Their pressure is idependent of events.

2. Automatic machines have to distinguish obstructive apneas - which justify a rise in pressure - from non-obstructives ones, that do not need raise in pressure.

3. The companies who make automatic machines going about this identification proscess in different ways.

Resmed's Auto Set Algorithm is decribed here: http://www.resmed.com/en-au/clinicians/ ... clinicians

And here: http://www.talkaboutsleep.com/sleep-dis ... chat.htm#6

The A10 algorithm increases pressure in response to Flow Limitation, Snore, and Apnea up to 10cm H2O. Above 10cm H2O, pressure response to Flow Limitation and Snore continues, but there is no response to Apnea. AutoSet Spirit and AutoSet T do not differentiate between obstructive and central apneas. Increasing pressures above 10cm H2O in response to apnea can lead to "runaway" central apneas,

This algorithm is still in use in the S8 machines.
If a person has apneas, but no snores or flow limitations above a 10 cwm, then a Resmed may not be the best machine for treatment. However, the majority of people have flow limitation and snores as well as apneas, and Resmed think it is safe enough to stop at 10.

Respironics idenfity apneas /hypopneas that do not respond to pressure, and then drop the pressure back down after having raised it too much.

Above 8 cmH2O pressure, the pressure increase for sustained apneas/hypopneas is limited to 3 cmH2O above the pressure setting at the onset of the apnea/hypopnea sequence. The pressure setting at the onset of the sequence is called the "Onset Pressure." The pressure 3 cmH2O above the "Onset Pressure" is called the "NRAH Threshold." "NRAH" is an acronym for "Non-Responsive Apnea/Hypopnea."
<snip>
When the device encounters a non-responsive apnea or hypopnea, it will decrease pressure by 2 cmH2O and hold the pressure for 15 minutes. During this 15-minute "NRAH-hold" period, the pressure will be changed only in response to detection of a sequence of vibratory snore events. The reason for this increase in pressure followed by a decrease in pressure when an apnea/hypopnea is not responsive is to allow the device to respond appropriately to an event that is not treatable by increases in pressure, such as a central apnea

.
Emphasis added by me.

Your machine is raising the pressure, Mike, because your apneas are responsive to those pressure raises. You do not not sustained apneas that are not responding the rise in pressure.

On Puritan Bennett automatic machines, you can tell the machine what the maximum pressure will be in response to apnea - the default is 10, but it can be set both higher or lower. The machine is also capable of identifying hearbeat oscillations in the air flow. Their existence idicates that the airway is open, and therefore the machine will not raise pressure in response to apneas accompanied by cardiogenic oscillations - no matter what pressure they occur at.

All of these algorithms will raise pressure as higher than 10 if they run into flow limitation or snores.

O.

Hi Ozij

Thanks so much for the clarification - very interesting...

So just to make sure I understand, you are saying that on the Remstar M-Series APAP, when in APAP mode, when an apnea event is detected the machine will respond by incrementally increasing the pressure to a maximum value of 3.0cm above the pressure where it was first detected - if the apnea event(s) is not resolved by increasing the pressure to a maximum value of (onset pressure +3.0), then it is considered to be a non-responsive event & the machine decreases the pressure again for a 15 min hold period... etc.

On a separate note, after 6 weeks, my typical AHI value is 6.0+ - given the data, I am trying to determine what I can do to get to the next level & reduce this value further - I am not sure whether my main issue is leaks ? I do see some leaks on the graph, but I am not really sure how significant they are in relation to the "intended leak" rate ?

Thanks
Mike

[Snoredog]

There have been many threads where the subject of "machines not responding to Apneas above 10cm" has been discussed.

Snoredog, I know you have clarified a number of times (at least in the case of the Remstar APAP machine) that the pressure will not increase above 10.0cm in order to respond to a Frank Apnea & that it will only go above 10.0cm to deal with Flow Limitation or Snoring events. (from what I have read, the only way around this is to increase the Min Auto setting to a value > 10.0)

Apologies, but I am still very confused on this and trying to figure out what my data represents - I have attached detailed results from 3 different days below - (my Min Auto setting is 10, my Max Auto setting is 15 & I am using a ramp that starts at 8.0) - in each case my indicated 90% AutoCPAP is between 11 and 15 cm, yet I see nothing/virtually nothing in the way of Flow Limitation or Snoring events to explain why the pressure is rising above 10.0, if it is not to deal with Apnea Events... what am I missing, or is my data screwed up because I still have leak issues to deal with ?

Thanks
Mike

Obviously it seems I didn't explain it in simple enough terms because you apparently still don't understand it. Where I have indicated that the machine does NOT respond at or above 10 cm was in the the case with the Resmed Auto CPAP machines.

Remstars DO respond above that pressure but use different command on apnea parameters. They respond and ask questions later by use of their NRAH algorithm.

[mymontreal]

I can promise you that it has sunk in this time Snoredog...

When you are sleep deprived, any machine starting with the letter "R" sounds the same...

Any thoughts/suggestions on what next steps I should look at in terms of trying to get that AHI value lower ?

Thanks
Mike

[Snoredog]

Any thoughts/suggestions on what next steps I should look at in terms of trying to get that AHI value lower ?

Thanks
Mike

My opinion is: you don't have a lot of leader type events leading to your Obstructive Apnea, that being vibratory snore (yeah a few snores seen but not many), Flow Limitation or Hypopnea. They seem to show up at or about 3 hrs into therapy, then some when seen are stubborn resulting in double apnea being shown.

This would indicate to me they are showing up during REM which averages in multiples of 90 minutes, so 3 hrs is about 2 REM cycles, appears to be where main clusters start. You may have 3 to 5 of those per night.

You had 1 or 2 periods where pressure pegged at 15 cm, therefore in the absence of CA, I would increase the Maximum to 20 so machine has the working pressure available to dislodge a stuck tongue if needed.

If some nights it seems you have more events or periods where they are absent, that could be from positional sleep changes. For example, when you are on your side very few if any apnea are seen, if on your back or supine, mainly more stubborn apnea are seen but machine appears to be handling it as there are no NRAH flags seen. My guess is most of your obstructions are from a tongue being sucked back into the throat, that would account for the apnea seen without the leader event.

I would go over your PSG and look for any indication that Central Apnea (CA) if none was seen, increase Maximum pressure to 20 cm. Mouth leak looks good and under control, use of Ramp has not impacted your therapy any, I would continue to use that, extend the time if you want, you can see where you used that, one report even shows you hit the Ramp at therapy hour 4 which is fine.

You might try increasing the Minimum by 1 cm to 11 cm, if you continue to use the Ramp you won't even notice the change but if it reduces the number of OA seen and the few residual snores that could only be a good thing.

[dsm]

1. The issue of responding after an apnea is identified is only relevant to automatic pressure machines. On fixed pressure machines, the pressure is fixed and if a person needs a pressure of 12 to keep them from having most of their apnea, that's what they'll have. Their pressure is idependent of events.

2. Automatic machines have to distinguish obstructive apneas - which justify a rise in pressure - from non-obstructives ones, that do not need raise in pressure.

3. The companies who make automatic machines going about this identification proscess in different ways.

Resmed's Auto Set Algorithm is decribed here: http://www.resmed.com/en-au/clinicians/ ... clinicians

And here: http://www.talkaboutsleep.com/sleep-dis ... chat.htm#6

The A10 algorithm increases pressure in response to Flow Limitation, Snore, and Apnea up to 10cm H2O. Above 10cm H2O, pressure response to Flow Limitation and Snore continues, but there is no response to Apnea. AutoSet Spirit and AutoSet T do not differentiate between obstructive and central apneas. Increasing pressures above 10cm H2O in response to apnea can lead to "runaway" central apneas,

This algorithm is still in use in the S8 machines.
If a person has apneas, but no snores or flow limitations above a 10 cwm, then a Resmed may not be the best machine for treatment. However, the majority of people have flow limitation and snores as well as apneas, and Resmed think it is safe enough to stop at 10.

Respironics idenfity apneas /hypopneas that do not respond to pressure, and then drop the pressure back down after having raised it too much.

Above 8 cmH2O pressure, the pressure increase for sustained apneas/hypopneas is limited to 3 cmH2O above the pressure setting at the onset of the apnea/hypopnea sequence. The pressure setting at the onset of the sequence is called the "Onset Pressure." The pressure 3 cmH2O above the "Onset Pressure" is called the "NRAH Threshold." "NRAH" is an acronym for "Non-Responsive Apnea/Hypopnea."
<snip>
When the device encounters a non-responsive apnea or hypopnea, it will decrease pressure by 2 cmH2O and hold the pressure for 15 minutes. During this 15-minute "NRAH-hold" period, the pressure will be changed only in response to detection of a sequence of vibratory snore events. The reason for this increase in pressure followed by a decrease in pressure when an apnea/hypopnea is not responsive is to allow the device to respond appropriately to an event that is not treatable by increases in pressure, such as a central apnea

.
Emphasis added by me.

Your machine is raising the pressure, Mike, because your apneas are responsive to those pressure raises. You do not not sustained apneas that are not responding the rise in pressure.

On Puritan Bennett automatic machines, you can tell the machine what the maximum pressure will be in response to apnea - the default is 10, but it can be set both higher or lower. The machine is also capable of identifying hearbeat oscillations in the air flow. Their existence idicates that the airway is open, and therefore the machine will not raise pressure in response to apneas accompanied by cardiogenic oscillations - no matter what pressure they occur at.

All of these algorithms will raise pressure as higher than 10 if they run into flow limitation or snores.

O.

Hi Ozij

Thanks so much for the clarification - very interesting...

So just to make sure I understand, you are saying that on the Remstar M-Series APAP, when in APAP mode, when an apnea event is detected the machine will respond by incrementally increasing the pressure to a maximum value of 3.0cm above the pressure where it was first detected - if the apnea event(s) is not resolved by increasing the pressure to a maximum value of (onset pressure +3.0), then it is considered to be a non-responsive event & the machine decreases the pressure again for a 15 min hold period... etc.

On a separate note, after 6 weeks, my typical AHI value is 6.0+ - given the data, I am trying to determine what I can do to get to the next level & reduce this value further - I am not sure whether my main issue is leaks ? I do see some leaks on the graph, but I am not really sure how significant they are in relation to the "intended leak" rate ?

Thanks
Mike

Mike,

The way you have worded your understanding can be misinterpreted. When an Auto responds to Apneas it is always neccessary to say it is responding to a group of apneas. It is very easy for necomers to misread the description as meaning that the Auto is responding to a single apnea which confuses the story. Put simply

1) When one apnea is detected the Auto starts a 20-30 second cycle of slowly raising the pressure by 1 or less CMs
2) The majority of apneas pass in less than 20 secs
3) If another apnea follows, the Auto will start another 20-30 sec cycle to increase by 1 or less CMs
4) The apnea will have passed before that cycle completes but if another apnea occurs the cycle at 1 is repeated again

After raising pressure by 3 CMs in response to a cluster of apneas. The machine will stop. raising pressure.

The purpose of slowly raising pressure is the expectation that the airway will be forced to open more & thus hopefully break the chain of apneas.
Put simply An AUTO does not clear a single apnea on the spot, it raises pressure far to slowly to do so. An AUTO clears a pattern of apneas.

When people understand that AUTOs raise pressure slowly over time, they begin to see that starting the AUTO at a low setting of say 4 is a recipe for having large clusters of apneas that do not get properly dealt with if their titration pressure was say 10 CMs. It takes a lot of AUTO response cycles to go from 4 CMs to 10 CMs if 10CMs is the best setting for the person on the machine.

Hope this helps

DSM

[Snoredog]

Mike,

I should add; Had you been using a "Resmed" Auto, it most likely would not have ever reached 14 cm pressure as shown in your graph. About the only thing it would have responded to seen in that top report was the single Flow Limitation. Since you have virtually no snore seen in your report there wouldn't have been anything for it to increase it.

However, the Resmed might have increased as I believe it is more sensitive to snore than the Remstar, this only from my own experience with that machine's algorithm as it doesn't keep track of and report snore even though it is one of the two things it responds to once over 10 cm. Flow Limitation is the other, it doesn't report FL's on the LCD or reports either.

At the same time, had the Resmed been reporting these events it might have seen a lot more FL's and Snores where NONE are shown on the Encore report, you just don't know.

I can use my Aflex and pull a Encore report and have over 200 snores, switch to my 420e and it shows very few snores. There is no real rhyme or reason to how they detect and count snores.

But after seeing your Encore reports, I would have to conclude the Resmed Auto wouldn't be a very good fit for you based upon your SDB event types.

[dsm]

Mike,

<snip>
However, the Resmed might have increased as I believe it is more sensitive to snore than the Remstar, this only from my own experience with that machine's algorithm as it doesn't keep track of and report snore even though it is one of the two things it responds to once over 10 cm. Flow Limitation is the other, it doesn't report FL's on the LCD or reports either.

<snip>

Snoredog,

You may not have realized this but the Resmed machines do record and report detailed snore data plus flattening info. But to get at that data I used a reslink. I think you are making some great leaps of faith in some of the things you are saying the Resmed machines don't do. The data shown below clearly shows otherwise

Here is an example off my S8 Vantage. (The S7s provide the same).
http://www.internetage.com/cpapdata/dsm ... 3jun08.pdf

DSM

#2 Just to make really clear - F/Ls, Hypops & Snores are all recorded & held in minute detail per the Snore data & flattening info, the MV data & pressure.

#3 - from the report ...

Minute Ventilation is the volume of air breathed in (or out) within any 60 second period.

Flattening Flattening is a measurement of partial upper airway obstruction. This
measurement is based on the shape of the inspiratory flow-time curve. A flat
shape suggests upper airway obstruction.

Snore Snore index is the measure of the amplitude of pressure wave generated by
a patient's snoring.

Leak is an estimate of the total rate of air escaping due to mouth and mask
leaks. It is derived by analyzing the inspiratory and expiratory airflows, together with
the expected mask vent flows. High or changing leak rates may affect the accuracy of
other measurements.

[Snoredog]

Snoredog,

You may not have realized this but the Resmed machines do record and report detailed snore data plus flattening info. But to get at that data I used a reslink. I think you are making some great leaps of faith in some of the things you are saying the Resmed machines don't do. The data shown below clearly shows otherwise

Here is an example off my S8 Vantage. (The S7s provide the same).
http://www.internetage.com/cpapdata/dsm ... 3jun08.pdf

DSM

I try NOT to mislead people. I don't like to mislead patients into thinking they are going to see or get that info from their S8 AutoSet as your reports are generated in another country using a clinical device they cannot legally purchase in the USA according to the manufacturer.

Resmed themselves say that Reslink module is a clinical only device, not to be sold to patients, neither is the software you printed it from if you want to get technical about it.

They could probably buy crack in another country too if they wanted.

[Snoredog]

Mike,

When Snoredog talks of machines not responding to apneas above 10CMs he is usually referring to the Resmed machines that use the A10 algorithm.

The machines do respond to apneas over 10CMs but that then turns into a debate of what apnea means. If we accept that there can be apneas that are 100% obstructive (no air flow) and that there are apneas that are called flow-limitations/hypopneas (some airflow) and there are apneas that are Central (no airflow) then the A10 algorithm can be correctly said to deal with some of these apneas, over 10CMs, but it does so selectively.

The A10 algorithm will regard certain of these apneas as potential Centrals and on that basis will reasonably seek leave them alone. Continually slowly adding pressure to a central apnea compounds it if the pressure keeps slowly rising (which is waht AUTOs do). If we look at how a timed bilevel reponds to a Central, we can see that it applies a burst of pressure (flips from epap to ipap) for a short period in the hope that that ipap burst jolts the sleeper back into breathing. Slowly increasing pressure doesn't clear a central but can disturb the sleeper if the pressure rise isn't stopped.

If A10 detects snores over 10 CMs it will rightly regard these as precursor events leading to a genuine obstructive apnea so it *does* raise its pressure & thus is responding to a *looming* obstructive apnea. If there are no snores and suddenly there is no flow (which could be a Central or an obstructive apnea) the A10 algorithm quite reasonably decides to leave it to nature. So, based on the situation that the A10 algorithm did not detect the typical obstructive apnea precursor events (snores & flow-lims/hypopneas) the correct statement is

The A10 algorithm won't respond to Obstructive apneas over 10 CMs *if* there are no obstructive apnea pre-cursor events that precede the obstruction. And, in reality there is nearly always some pre-cursor obstructive apnea events that precede obstructive apneas.

I believe myself, that it is a mis-statement to say that 'the A10 algorithm doesn't respond to 'apneas' over 10 CMs' without spelling out clearly what is meant by apnea- because not clarifying that meaning twists the reality.

DSM

I think this is your "opinion" more so than actual fact.

I'm willing to sit here while you explain how it responds to apnea at or above 10 cm. This I gotta see.

Below is what it says about the AutoSet from the Clinical manual (figures missing):

The treatment pressure required by your patient may vary through the night, and
from night to night, due to changes in sleep state, body position, and airway
resistance. In AutoSet mode AUTOSET SPIRIT provides only sufficient pressure to
maintain upper airway patency.
You can set the minimum and maximum allowable treatment pressures. The unit
analyzes the state of the patient’s upper airway on a breath-by-breath basis and
delivers pressure within the allowed range according to the degree of obstruction.

The AutoSet algorithm adjusts treatment pressure as a function of three parameters:
inspiratory flow limitation, snore, and apnea.

The flow sensor, located in the AUTOSET SPIRIT unit, enables detection of inspiratory
flow limitation and apneas. The pressure sensor, also located in the unit, enables
measurement of pressure and snore.
Inspiratory flow limitation indicates silent partial obstruction. When your
patient is breathing normally, the inspiratory flow measured by the unit as a
function of time shows a typically rounded curve for each breath.

The AutoSet algorithm analyzes the shape of the central part of the curve for each
breath and assigns a value as the amount of flattening.
Inspiratory flow limitation, or partial airway closure, usually precedes snoring and
obstruction. Detection of this flow limitation enables the unit to increase the
pressure before obstruction occurs, making treatment pre-emptive.

The Flattening Index is a measure of silent inspiratory airflow limitation. Flow
limitation with loud snoring is handled by the snore detector. When a patient
snores, sound is generated and the inspiratory flow/time curve is distorted by the
frequency of the sound.

The AutoSet algorithm assigns an arbitrary value between 0.0 and 2.0 to the
average amplitude of the snoring detected for the past 5 breaths. A value of 1.0 is
equivalent to approximately 75dBA measured 10cm from the nares. Treatment
pressure increases by up to 0.2 cm H2O per second (proportional to the severity
of the snore) for snore above 0.2 snore units. When snore is less than 0.2 snore
units, therapy is reduced towards the minimum pressure with a 20-minute time
constant.

An apnea is defined as a greater than 75% decrease in ventilation. The AutoSet
algorithm scores an apnea if the 2-second moving average ventilation drops below
25% of the recent time average (time constant 100 seconds) for at least 10
consecutive seconds. Treatment pressure increases based on the duration of the
apnea. The pressure will not rise above 10 cm H2O when an apnea is detected, to
prevent an inappropriate response to central apneas. Initial pressure increases are
rapid, but the rate of increase diminishes as the pressure approaches 10cm H2O.
When no further apneas are detected, therapy is reduced towards the minimum
pressure with a 20-minute time constant.

A hypopnea is defined as a 50 to 75% drop in ventilation. A hypopnea is scored if the
8-second moving average ventilation drops below 50%, but not below 25%, of the
recent average for 15 consecutive seconds. In order to avoid falsely responding to
central hypopneas, the AutoSet algorithm does not respond to hypopneas but rather
to the associated snore or flow limitation.

This is all kind of mute in Mike's case, if your pressure to keep your airway patent is over 10 cm like Mikes is, and in absence of FL or Snore the machine isn't going to respond just like it says in the above paragraph. He might as well be on CPAP treatment because the Minimum pressure would have to be set the same as CPAP in order to take care of his apnea. IF he snored or had FL above 10 cm then the Resmed would respond.